Cornell researchers are taking a new approach to understanding how voters’ attitudes about immigration, race and gender influence contemporary U.S. politics.

An innovative study using three waves of surveys will show how voters’ views on those topics influence the midterm elections in November and whether those attitudes shift leading up to the elections.

“We’re especially interested in race, immigration and gender attitudes and the midterm elections, but we’re also interested in the broader political environment, including the relationship between these issues and voters’ support for President Donald Trump,” said Peter K. Enns, associate professor of government and executive director of the Roper Center for Public Opinion Research at Cornell. Enns and Jonathon P. Schuldt, associate professor of communication and a faculty affiliate at the Roper Center, are working with Adrienne Scott, a Cornell Ph.D. student in government, on the project.

The 2018 midterm election study, funded by Cornell’s Center for the Study of Inequality, is surveying a nationally representative sample of nearly 1,400 likely voters of all political persuasions. The first survey was conducted in early July; subsequent surveys will take place in late August or early September, and again in October. The results are just in from the first survey and are available at the Roper Center. Researchers also will make data from the subsequent two surveys available at the Roper Center.

The study is unusual in several ways.

Its three-wave structure means the same individuals are interviewed at three points prior to the midterm election in November. Some questions will be repeated every survey, some will appear in the first and last, and others will be unique to each survey. The questions asked in the first and last survey are especially important, the researchers say.

“Repeating questions in July and October means that instead of simply observing whether or not racial bias or immigration attitudes are correlated with vote choice, we can observe if a change in bias corresponds with a change in vote intention,” Enns said. Schuldt added, “As the congressional campaigns unfold, we will actually be able to observe what factors correspond with shifts in vote choice.”

Some questions are designed to help reveal racial attitudes, such as “Do you support or oppose the movement called Black Lives Matter?” and “[To what degree do you agree or disagree that] many groups in the U.S., like Irish, Italian and Jewish people, have overcome prejudice and worked their way up, and blacks should do the same without any special favors?” The first of the three surveys found that 89 percent of those who strongly support Black Lives Matter indicate they would vote for the Democratic candidate. Among those who strongly oppose the movement, just 8 percent indicate a Democratic vote intention. Similar patterns emerge with the question on prejudice toward immigrant groups.

The study also is attempting to untangle racial bias against African-Americans and anti-immigrant bias, which the media and academics have routinely lumped together in the past few decades, Enns said. “Measuring those two biases through separate questions and allowing them to be separate concepts is a really important part of the survey research design.”

The July survey asked questions on immigration, race, gender, political leanings and attitudes about Trump.

In terms of immigration, most respondents – 35 percent – said it’s not very likely that recent immigration will take jobs away from people already in the U.S. Additional findings about immigration were featured in Political Wire.

When asked how they would vote if the election for the U.S. House of Representatives were held today, 51 percent said they would vote Democrat, 40 percent said Republican, and 9 percent said they would vote for neither party. When asked how they would vote if the election for the U.S. Senate were held today, 52 percent said they would vote Democrat, approximately 40 percent said Republican, and 7 percent said they would vote for neither party.

The researchers used a standard survey question to separate Democrats and Republicans into different groups: those who say they are independent but “lean” toward their respective parties, those who identify with their parties, and those who strongly identify with their parties. They then combined this question with a unique approach that allowed them to determine how likely voters rank Donald Trump relative to other Republican politicians.

The survey found strong Republicans rank Trump over other Republican leaders including former Presidents George W. Bush and Ronald Reagan. However, despite conventional wisdom, the survey showed that Trump’s political base is weaker than it seems. That’s because many media reports don’t factor in the differing levels of party affiliation, Enns said.

“… Portraying Trump’s support among Republican voters as unflinching is missing a major piece of the picture,” the study’s co-authors wrote in an Aug. 7 Washington Post opinion piece.

The team will continue to analyze implications of the results of the surveys over the next two years, they said.